[Orca-users] Question regarding Statistical Calculations...

[Sadat Chowdhury]

Jon,

Thanks a lot for your quick reply. Do you know approximately when you 
might be done with modifications needed? In the meantime, if you can 
tell me which particular (perl) modules to tweak, I might just play 
around with it myself...

Again, I really appreciated your quick response.

-Sadat.

Jon Tankersley wrote:

>This is a result of the rrdtool consolidation function that is used by
>default by Orca - which is AVERAGE, so weekly is averaged again and drops
>the peak value, as is monthly, yearly, etc.  We ran into this and have some
>'hooks' that I'm trying to put together for Blair to allow you to specify in
>the config line (instead of changing the source) how you want data to be
>consolidated - currently with min, max, avg separately or all together or
>max and avg together (our current default usage).  You then specify in the
>plot section how you want to display the data.  I'm hoping to post it to
>Blair fairly soon.
>
>This also includes a change in the legend to make it clear that you are
>using an average, min or max:
> % Usr        Curr Max:   20.091   Avg:  35.025   Min:  15.099   Max:
>60.092
> % Usr        Curr Avg:   19.991   Avg:  34.714   Min:  15.099   Max:
>64.012
> % Sys        Curr Max:    5.312   Avg:   8.127   Min:   2.091   Max:
>12.209
> % Sys        Curr Avg:    5.102   Avg:   8.018   Min:   2.091   Max:
>11.001
>
>Here, the Curr XXX is the last data point (on the right of the graph) and
>the other fields are the Avg, Min, and Max of the entire interval.
>With this setup, the Weekly, Monthly, and Yearly graphs all have a retained
>real peak.
>
>
>-----Original Message-----
>From: orca-users-bounces+jon.tankersley=eds.com at orcaware.com
>[mailto:orca-users-bounces+jon.tankersley=eds.com at orcaware.com] On Behalf Of
>Sadat Chowdhury
>Sent: Wednesday, July 14, 2004 12:29 PM
>To: orca-users at orcaware.com
>Subject: [Orca-users] Question regarding Statistical Calculations...
>
>
>
>We have been using the orca system on a solaris machine and so far we 
>have been able to get the system to gather data and generate html/graphs.
>
>The statistics from the graphs, however, look a bit confusing. Here are 
>some values quoted:
>
>Daily:
>Disk...Max: 3.56
>1 min avg Process in Run Queue: Max: 5.753
>5 min avg Process in Run Queue: Max: 4.990
>15 min avg Process in Run Queue: Max: 4.150
>
>Weekly:
>Disk....Max: 2.13
>1 min avg Process in Run Queue: Max: 3.901
>5 min avg Process in Run Queue: Max: 3.966
>15 min avg Process in Run Queue: Max: 3.966
>
>The statistics generated for daily and weekly are from the same data 
>set. They were both generated at the same time. How come the weekly max 
>values are LOWER than the daily max values, when in fact, one would 
>expect them to be at least the same or higher? For instance, if on 
>Monday the max is 5.753 for any measure, for the entire week, it can't 
>possibly be 3.901 - it has to be at least 5.753 or more (say of the max 
>on Friday was 7.901).
>
>Is there a reasonable explanation? anyone else experienced this?
>
>-Sadat.
>
>_______________________________________________
>Orca-users mailing list
>Orca-users at orcaware.com http://www.orcaware.com/mailman/listinfo/orca-users
>
>  
>